Acute Respiratory Distress Syndrome (ARDS) and acute lung injury (ALl) place heavy burdens on the health of the public, afflicting over 190,000 people annually in the US alone. The associated mortality exceeds 30% and accounts for 74,500 deaths annually. Improvements in ventilator management have reduced both the mortality and morbidity associated with ALI/ARDS, and thus have an important public health impact. Low tidal volume ventilation strategies are clearly beneficial in patients with ALI/ARDS. However, the optimal level of applied positive end-expiratory pressure (PEEP) is uncertain. We propose a phase II multicenter, randomized controlled trial of mechanical ventilation directed by esophageal pressure measurements. The primary hypothesis is that adjusting PEEP to maintain a positive transpulmonary pressure throughout the ventilatory cycle will lead to clinically important improvements at 28 days. In patients with high pleural pressure on conventional ventilator settings, underinflation leads to atelectasis, hypoxemia, and exacerbation of lung injury through atelectrauma. In such patients, raising PEEP to maintain a positive transpulmonary pressure might improve aeration and oxygenation without causing overdistension. Conversely, in patients with low pleural pressure, maintaining a low PEEP would keep transpulmonary pressure low, avoiding over-distension and consequent volutrauma. Thus, the currently recommended strategy of setting PEEP without regard to transpulmonary pressure could benefit some patients while harming others. Our preliminary study demonstrated that a ventilator strategy designed to optimize transpulmonary pressure significantly improved oxygenation and lung mechanics, and there was a trend toward improved survival. We propose to confirm these results in a multi-centered trial at 6 large medical centers enrolling 200 patient with ALI/ARDS. We will use esophageal pressure to estimate transpulmonary pressure (PTP). Subjects will then be randomized to the control group, in which ventilation will be managed according to the ARDSnet protocol, or the EPVent group, in which ventilator settings will be adjusted to achieve a specified PTP. The primary hypothesis is that using a strategy of maintaining a positive PTP throughout the ventilatory cycle will lead to improvement in a combined hierarchical outcome of mortality and ventilator free days at 28 days. This study has important implications for the health of the public. Strong preliminary data suggest a positive result, potentially changing medical practice.